The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) and official GitHub APIs (REST and GraphQL) to automate the creation and management of issues and labels. These operations are confined to the repository's issue tracking system and are standard for development workflow tools.
[PROMPT_INJECTION]: The skill processes data from GitHub issue bodies (Ingestion point: Step 1, SKILL.md). Although this content is externally controlled, the risk of indirect prompt injection is mitigated by a mandatory human-in-the-loop checkpoint in Step 6, where the user must approve the generated task breakdown before any issues are created (Capability: Step 8). No specific sanitization is mentioned, but the structural template in references/issue-template.md acts as a boundary marker for the generated output.
[DATA_EXPOSURE]: The skill performs read-only exploration of the codebase and issue metadata to inform the task splitting process. All data access is performed through standard interfaces within the agent's authorized scope.

ito-issues